Amazon is hosting AWS Re:INVENT today, its developer conference for all things AWS. And the company just announced a promising new service called Amazon Transcribe. The service is now available as a preview and goes beyond many automated transcribing services.
Now that video and audio have overtaken the web, it has become harder to parse information inside those media formats. One way to do it is to transcribe the audio part and turn it into text. Text is indexable, searchable and opens new possibilities.
With Amazon Transcribe, the company has built a speech recognition engine. It lets you turns an audio file stored on your Amazon S3 account into grammatically correct text.
Amazon Transcribe works in English and Spanish for now. But the company promised that many more languages will be added in the coming weeks.
The secret sauce behind Amazon Transcribe is that the service can intelligently format and add punctuation. The service can also recognize multiple speakers and adds timestamps so that you can isolate each part of a conversation.
Amazon mentioned multiple use cases. For instance, Transcribe can help you create automatic subtitles for online videos. It’s also a good way to log customer support calls and analyze them.
The service works with low-bitrate audio files, such as call recordings, and you can add your own vocabulary if you want to help the service understands product names.
Amazon Transcribe will also be useful with Amazon Translate and Amazon Comprehend, two new services announced today. This way, you can turn audio recordings into meaningful data.